autopkgtest failure with python-biopython 1.79+dfsg-1~0exp0 in experimental

Source: ncbi-acc-download
Version: 0.2.7-1
Severity: important
Tags: ftbfs

Dear Maintainer,

I am trying to assess the side effects of an upgrade of the
package python-biopython to 1.79 to its reverse dependencies.
Pseudo-excuses look alright, except for ncbi-acc-download [1].


The full log [2] shows variations around the following symptom:

        __________________ test_download_wgs_parts_supercontig_retry 
        req = <requests_mock.mocker.Mocker object at 0x7fc14613fb20>
            def test_download_wgs_parts_supercontig_retry(req):
                cfg = Config(format="genbank")
                supercontig = open(full_path('supercontig.gbk'), 'rt')
                req.get(ENTREZ_URL, response_list=[
                    {"text": u'Whoa, slow down', "status_code": 429, "headers": 
{"Retry-After": "0"}},
                    {"body": open(full_path('supercontig_full.gbk'), 'rt')}
                outhandle = wgs.download_wgs_parts(supercontig, cfg)
                supercontig_full = open(full_path('supercontig_full.gbk'), 'rt')
        >       assert outhandle.getvalue() ==
        E       AssertionError: assert 'LOCUS       ...671808)n//n' == 'LOCUS 
        E         - LOCUS       NC_007194                 60 bp    DNA     
linear   CON 03-APR-2018
        E         ?                                  ^^^^^^^                    
        ^^ ^^^   ^^
        E         + LOCUS       NC_007194            4918979 bp    DNA     
linear   CON 11-NOV-2009
        E         ?                                  ^^^^^^^                    
        ^^ ^^^   ^^
        E           DEFINITION  Aspergillus fumigatus Af293 chromosome 1, whole 
genome shotgun
        E                       sequence.
        E           ACCESSION   NC_007194...
        E         ...Full output truncated (22 lines hidden), use '-vv' to show

To be honest, I don't know whether this is caused by a bug in
biopython, or if it could be normal from ncbi-acc-download to
issue a slightly different output.  Testing against biopython
1.78 does not raise that issue.

Since the test suite is run at build time, ncbi-acc-download
would also fail to build from source if python-biopython were to
be upgraded, hence the tag.  Note that I think the package
downloads resources from the Internet during the test at build
time too, so would might be a serious policy violation, but I
haven't checked for sure yet.

Kind Regards,
Étienne Mollier <>
Fingerprint:  8f91 b227 c7d6 f2b1 948c  8236 793c f67e 8f0d 11da
Sent from /dev/pts/4, please excuse my verbosity.


Description: PGP signature

Read more here: Source link